On 12/16/2009 04:25 PM, John Plevyak wrote:
I wouldn't mind doing the April/May release because of the things which are not going to make it into this release: the cache partition stuff, large document support, efficient Range support, the event system work for libev/libevent.
But, what are the odds that we can do two such large releases in 2-3 months? I'm open to the idea, but also trying to be realistic. I'd imagine once we get 2.0 out, we'll have to spend a bunch of time doing bug fixes as people start using it. We still have the option to make a "2.1" release any time we like, which brave people can pick up (but which distros would be discouraged from touching).
Also, I definitely missed it, but I didn't know we had already decided about April / May and October / November release schedules. Since that's been brought up, we should make the "official" schedules be Q2/Q4 instead of what I said earlier, with the exception that we want an earlier first release in Q1 of 2010.
Basically the current code is unusable for anything of even moderate size.
Agreed. -- Leif
